Ali  Bashirpour, Fereydon Rahnamay  Roodposhti, Kazem  Chavoshi, Ebrahim  Saber, 
    Optimization of the Mutual-Fund Portfolio of Tehran Stock Exchange Using Artificial Neural Networks and Genetic Algorithm,
    International Business Management,
    Volume 10,Issue 11,
    2016,
    Pages 2249-2256,
    ISSN 1993-5250,
    ibm.2016.2249.2256,
    (https://makhillpublications.co/view-article.php?doi=ibm.2016.2249.2256)
    Abstract: The purpose of present research is to predict the return of mutual funds listed in Tehran Stock Exchange by a linear panel data regression model and a nonlinear model based on artificial neural networks then optimizing the portfolio consisting of these mutualfunds using Genetic Algorithm (GA). Using 13 factors affecting mutual fund returns as input data to both pooled regression and ANNs results showed the predictability of mutual fund returns by both methods using these factors but ANNs had a better performance. The results showed that GA can be used for mutual-fund portfolio selection and the superiority of GA method on Markowitz approach. Portfolio size had no significant effect on the results and GA had a better performance at all levels. GA had a markedly better performance than linear methods when portfolios were more diversified. So, investors can use nonlinear models such as ANNs to predict mutual-fund returns. Also investors can use nonlinear methods such as GA to build optimal portfolios. This research provides new insights into mutual funds and their driving factors in Iran’s emerging market. It is also a new approach to using nonlinear ANNs for prediction of mutual-fund returns and using GA for portfolio optimization.
    Keywords: Mutual fund Neural Networks (NNs);Genetic Algorithm (GA);portfolio optimization;panel data;factors